The Report Report presents a whitepaper on the race to physical AI.

AI has become remarkably good at understanding language, images, and code. The next challenge is harder — understanding reality, according to Analog Devices Inc. or ADI. Physical AI moves intelligence out of digital environments and into machines that must sense, decide, and act in the world.

Unlike software, those systems face imperfect measurements, timing constraints, changing conditions, and the consequences of action, said the company in a new report. Success will depend on more than model performance. It will require advances across sensing, simulation, computing, connectivity, and control.

In many cases, the technologies that determine whether a system succeeds sit beneath the application layer itself. A machine that shuts down when vibration crosses a preset limit is following a rule. A machine that reads vibration, temperature, current, and motor behavior; recognizes the signature of bearing wear; and adjusts before failure occurs is doing something closer to judgment.

That distinction helps explain why physical AI has become one of the most closely watched areas in technology for manufacturing. These systems do more than analyze information. They interact with the world around them, adapting to changing conditions and responding to situations that cannot be fully predicted in advance.

The next phase of physical and embodied AI is not simply about producing better answers. It is about predicting and producing useful actions in the presence of imperfect information, said ADI. To read the full whitepaper, click here: The Race to Make AI Physical .

Source link